Artificial intelligence-assisted smart robotic pipeline health monitoring system

近年来，加拿大和印度的管道天然气运输迅速增加。仅在印度，截至2021年底，已安装超过13，840公里的天然气管道，到2030年底，这一数字将达到28，000公里。作为世界第六大天然气生产国，加拿大的管道一直是长距离天然气的主要驱动力。然而，这些管道由于管道表面退化和接头失效的自然过程而具有有限的寿命。基于传感的结构健康监测（SHM）在提高管道寿命和预防灾难性故障方面具有巨大潜力。然而，目前可用的在线扫描机器人使用单一的传感方法，并不适合在管道中的大范围的实时泄漏检测。在单个机器人上创建磁性，触觉和红外扫描系统的组合，通过验证和融合来自所有传感器的数据来进行损伤预测，从而增强了更深入的学习。与传统的检测工具相比，这些由印度理工学院坎普尔（IITK）开发的独特传感器的扫描数据为管道的健康状况提供了宝贵的见解。然而，在没有任何自动化能力的情况下，该系统不足以充分利用数据融合并进行管道中的实时泄漏检测。该提案旨在将机器人传感设备的优势与人工智能相结合，并开发一种实时多级管道健康监测（PHM）系统，该系统可扩展到任何管道网络。在这项提案中，西方大学的Ayan Sadhu博士凭借其在人工智能和SHM方面的专业知识，将与IIT坎普尔的Bishakh Bhattacharya博士合作，后者开发了一种独特的机器人系统和测试平台来进行这项研究。加拿大主要管道机构（例如，Exxon、Enbridge、TransCanada等）将使用拟议的PHM系统，以提高管网的寿命，从而减轻加拿大的经济和环境损失。拟议中的系统将有助于减少由于管道故障而造成全球灾难的可能性。

Transportation of natural gas by pipelines has been rapidly increasing both in Canada and India in recent years. In India only, at the end of 2021, over 13,840 km of natural gas pipelines have been installed, and this number will be 28,000 km by the end of 2030. Being the sixth largest producer of natural gas in the world, pipelines in Canada have been the main driver of natural gas over long distances. However, these pipelines have limited life due to natural processes of degradation of the pipe surface and joint failures. Sensing-based structural health monitoring (SHM) has a significant potential to improve the life span of pipelines and prevent catastrophic failures. However, currently available in-line scanning robots use a single sensing approach and are not suitable for a wide range of real-time leak detection in pipelines. Creating a combination of magnetic, haptic, and infra-red scanning systems on a single robot enhances deeper learning by validating and fusing data from all sensors for damage prognosis. The scanning data from these unique sensors developed by the Indian Institute of Technology Kanpur (IITK) India provides valuable insight into the pipeline's health compared to the conventional inspection tool. However, in the absence of any automation capability, this system is inadequate to fully utilize the data fusion and undertake real-time leak detection in pipelines. This proposal aims to integrate the strength of robotic sensing devices with artificial intelligence and develop a real-time multiclass pipeline health monitoring (PHM) system, which can be scalable to any pipeline network. In this proposal, Dr. Ayan Sadhu at Western University, with his expertise in AI and SHM, will collaborate with Dr. Bishakh Bhattacharya at IIT Kanpur, who has developed a unique robotic system and testbed to conduct this research. Major Canadian pipeline agencies (e.g., Exxon, Enbridge, TransCanada etc.) will use the proposed PHM system to enhance the lifespan of the pipeline network and thus mitigate the economic and environmental loss in Canada. The proposed system will help reduce the possibility of worldwide catastrophe due to pipeline failure.